CPU switching from 10% to over 100%. - Page 2

JamesPeters · 12-09-2017, 07:33 PM

What plugins are on that first track listed in the performance meter, which require 27,648 samples of plugin delay compensation? That's enormous. Ok I see some are UAD plugins...here's from the user guide:

"Anticipative FX processing. This can reduce CPU usage, but too aggressive a setting can cause
pops and crackles on playback. If working with UAD-1,for best results, both Synchronous FX and UAD-
1 Synchronous modes should be enabled. Anticipative FX should be disabled. UAD-2, however, is
compatible with Anticipative FX mode and this is recommended for best low latency performance.."

I also notice you have been glossing over what some people have been asking about PDC etc. If you want to get help you need to provide info. We're not there in the room with you.

serr · 12-09-2017, 08:55 PM

Oh yeah look at that!

I'd try splitting those plugins across two (or more) tracks and see if that helps. Just daisy-chain route from one track to the next and split the plugins between them. I've had this come up before with large PDC. Splitting the plugins across multiple tracks gives you smaller PDC on any one track and this seems more important than the plugin's processing demands for system performance.

giankam · 12-10-2017, 04:29 AM

Great!!! I disabled the PDC in my main group track and I see no more spikes and the cpu is stable at a very low levels. Have I found the cure? How is this possible? What’s the drawback disabling the pdc?

Quote:

Originally Posted by serr

Oh yeah look at that!

I'd try splitting those plugins across two (or more) tracks and see if that helps. Just daisy-chain route from one track to the next and split the plugins between them. I've had this come up before with large PDC. Splitting the plugins across multiple tracks gives you smaller PDC on any one track and this seems more important than the plugin's processing demands for system performance.

serr · 12-10-2017, 10:33 AM

Quote:

Originally Posted by giankam

What’s the drawback disabling the pdc?

One would be if you were doing parallel processing work with a source split to multiple tracks. You would get an offset between tracks and have phasing issues.

If you are only monitoring the output of a single bus track, any lag is a moot point.

JamesPeters · 12-10-2017, 10:52 AM

Disabling PDC is what you did, not disabling anticipative FX?

See page 115 in the User Guide.

Normally disabling PDC isn't a big deal, but considering your PDC was 27,648 samples (divided by the samplerate 96,000 samples/sec) that means your plugin delay compensation was 288ms. That's a very noticeable delay, so this will mess up the timing of your tracks relative to one another.

I'd try to find another solution (such as setting the track to "prevent anticipative fx" under track performance options, found by right-clicking the track). If you haven't tried that already that is.

giankam · 12-10-2017, 11:23 AM

Thank you James, I'll try this out.

I've tried all the different options in Reaper preferences including disabling the global anticipative FX in the last few days but it didn't work out. As soon as I'm back to the studio I'll try these track specific options.

Quote:

Originally Posted by JamesPeters

Disabling PDC is what you did, not disabling anticipative FX?

See page 115 in the User Guide.

Normally disabling PDC isn't a big deal, but considering your PDC was 27,648 samples (divided by the samplerate 96,000 samples/sec) that means your plugin delay compensation was 288ms. That's a very noticeable delay, so this will mess up the timing of your tracks relative to one another.

I'd try to find another solution (such as setting the track to "prevent anticipative fx" under track performance options, found by right-clicking the track). If you haven't tried that already that is.

JamesPeters · 12-10-2017, 12:14 PM

Oh if you already tried disabling anticipative fx globally, I don't think disabling it per track will help.

I just realized there's something unusual looking about your sends and receives for the tracks in the pictures you attached. The folder track looks normal (master send), but track "audio 1" has a send and track "audio 2" has a receive (while both also have master sends). If you're sending from audio 1 to audio 2, and I'm assuming the audio is also going to the parent folder, that might have something to do with it (based on how the track "audio 2" has its own audio, and how the effects are set up possibly).

I actually don't see why you'd send from audio 1 to audio 2 if audio 2 has its own audio files on the track. Although, I don't know the specific details of what you're doing since I can only see what's in the pictures. But something seems weird about that arrangement. What sort of send/receive relationship are you going for?

Judders · 12-11-2017, 08:52 AM

Apologies if this has been mentioned, but I didn't see it skimming the thread:

Look at the bottom option. Seems relevant?

I don't have UAD, but noticed this today.

giankam · 12-12-2017, 03:51 AM

Hi James,

there's a side chain send from audio 1 to audio 2. The plug Trackspacer received the signal from this side chain. this explains the send. there's no send toward the parent folder though.

I tried to disable the single tracks anticipative FX option but it made things worse.

Quote:

Originally Posted by JamesPeters

Oh if you already tried disabling anticipative fx globally, I don't think disabling it per track will help.

I just realized there's something unusual looking about your sends and receives for the tracks in the pictures you attached. The folder track looks normal (master send), but track "audio 1" has a send and track "audio 2" has a receive (while both also have master sends). If you're sending from audio 1 to audio 2, and I'm assuming the audio is also going to the parent folder, that might have something to do with it (based on how the track "audio 2" has its own audio, and how the effects are set up possibly).

I actually don't see why you'd send from audio 1 to audio 2 if audio 2 has its own audio files on the track. Although, I don't know the specific details of what you're doing since I can only see what's in the pictures. But something seems weird about that arrangement. What sort of send/receive relationship are you going for?

giankam · 12-12-2017, 04:06 AM

Thanks for the suggestion but it looks that option works with the old UDA1 only.
I use the UAD2 which shouldn't be affected but I will try it anyhow.

Quote:

Originally Posted by Judders

Apologies if this has been mentioned, but I didn't see it skimming the thread:

Look at the bottom option. Seems relevant?

I don't have UAD, but noticed this today.

JamesPeters · 12-16-2017, 05:45 PM

I notice your thread priority and behavior settings have been altered. Perhaps setting them to that aggressive a setting is not a good match for something like the UAD card (which might need some priority left for its own software or driver?)

Fabian · 12-19-2017, 03:42 AM

Quote:

Originally Posted by serr

Huh... OK sorry, learned a new thing today then!

That article talks about it being a bug for Pentium CPUs in Windows machines circa 2003. Is that beside the point and this is a thing for any/all digital audio handling on any OS/system? Or an odd old Windows bug? (Sorry again, really not trying to start a Mac/Win thing.)

Maybe I need to start watching Activity Monitor a little more closely!

It is not a Pentium bug, it is a real problem with some plugins. It exists. It is not me who is the fool..

I have this problem with my AMD Phenom x6 1055T, with an old version of Boogex. Put GNormal in front of it and everything works fine, disable GNormal and the CPU % goes through the roof. If you want, I can share that "April's fool" project with you.